city auditor's office patrol operationsoperations+-+report+1703.pdf · the city auditor...

28
CITY AUDITOR'S OFFICE Patrol Operations June 16, 2017 AUDIT REPORT NO. 1703 CITY COUNCIL Mayor W.J. “Jim” Lane Vice Mayor Suzanne Klapp Virginia Korte Kathy Littlefield Linda Milhaven Guy Phillips David N. Smith

Upload: buicong

Post on 09-Sep-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

CITY AUDITOR'S OFFICE

Patrol Operations June 16, 2017 AUDIT REPORT NO. 1703

CITY COUNCIL Mayor W.J. “Jim” Lane Vice Mayor Suzanne Klapp Virginia Korte Kathy Littlefield Linda Milhaven Guy Phillips David N. Smith

TABLE OF CONTENTS

AUDIT HIGHLIGHTS ......................................................................................... 1

BACKGROUND ............................................................................................... 3

Figure 1. Patrol District Map ...................................................................................... 4

Figure 2. Scottsdale Police Department’s Dispatch Priorities ............................................... 5

Figure 3. Calls for Service (CFS) Process ........................................................................ 6

OBJECTIVES, SCOPE, AND METHODOLOGY .............................................................. 7

FINDINGS AND ANALYSIS ................................................................................... 9

1. The Patrol position reductions have had limited effect on the operational measures that were previously reported by CPSM. ................................................................ 9

Patrol position reductions have not resulted in significant changes in response times. ...... 9

Table 1. Average Response Times in Minutes: Citizen-generated Calls for Service ................... 10

Average occupied time is now slightly higher than in the CPSM study. ......................... 11

Table 2. Average Occupied Time per Call in Minutes, by Call Type and Source ....................... 11

On average, fewer Patrol units are now responding to calls than previously reported. ..... 12

Table 3. Average Number of Responding Patrol Units, by Call Type and Source ...................... 13

Figure 4. Number of Responding Patrol Units, by Percentage of Calls For Service .................... 14

Figure 5. Average Number of Responding Patrol Units, by Call Type .................................... 15

Patrol is spending more time on reactive calls, but specialty units augment Patrol’s proactive time. ....................................................................................... 16

Figure 6. Patrol Time Spent on Calls for Service and Other Activities in Selected Periods .......... 17

2. Procedural changes can better assure reliable data for operational analysis............... 18

MANAGEMENT ACTION PLAN ............................................................................. 21

Page 1

An audit of Patrol Operations was included on the City Council-approved fiscal year (FY) 2016/17 Audit Plan. The audit objective was to evaluate the effect of the FY 2016/17 staffing reorganization on previous CPSM study results for Patrol Operations, such as comparative workloads, response times and time spent on calls. In May 2015, the former City Manager contracted for the Center for Public Safety Management (CPSM) to review operations of the Scottsdale Police Department. The final report, delivered in January 2016, included certain operational statistics and compared these results to other police departments that CPSM had reviewed. With the FY 2016/17 budget request, the Police Chief proposed a reorganization plan reducing ten sworn police officer positions and adding 3.6 full-time equivalent positions in other areas. The reorganization plan reduced Patrol staffing by 13 officers: 9 were reduced through attrition, 3 were moved to the Bike Unit, and 1 was moved to the Training Unit. The proposed reorganization plan was approved and Patrol deployment was adjusted in July 2016. This audit analyzes the department’s FY 2016/17 dispatch data through March 2017 to calculate operational statistics for comparison to the FY 2014/15 CPSM-reported statistics.

AUDIT HIGHLIGHTS

WHY WE DID THIS AUDIT

BACKGROUND

Patrol Operations June 16, 2017 Audit Report No. 1703

WHAT WE FOUND

The Patrol position reductions have had limited effect on the operational measures that were previously reported by CPSM. • Patrol position reductions have not resulted in significant changes in

response times. Average response times for high priority calls (priorities 0, 1, or 2) increased only slightly, from 6.10 minutes to 6.18 minutes, or by about 1.3%.

• Average occupied time per call is now slightly higher than in the CPSM study. For citizen-generated Calls for Service (CFS), average occupied time increased from about 43.8 minutes to about 49 minutes, or by about 11.8%. Overall average time spent on officer-initiated calls increased only slightly from 27.9 to 28.6 minutes.

• Fewer Patrol units are now responding to calls than previously reported. The average number of units responding to citizen-generated CFS is 1.7, lower than the 1.9 average reported for FY 2014/15.

• Patrol units are spending more of their time on reactive calls than previously. However, other Uniformed Services specialty units, such as Bike, Motor and DUI, supplement Patrol’s reduced proactive time.

Procedural changes can better assure reliable data for operational analysis. • Testing entries in the live CAD system were not consistently tracked and

monitored.

• The department’s internal response time report excludes certain call types or events which can distort the results.

WHAT WE RECOMMEND We recommend the Police Chief:

• Direct the department’s Technology group to develop a detailed procedure for any testing conducted in a live system, such as requiring test entries to be consistently identified, monitored and removed, if possible.

• Ensure the internal response time report is consistent with the department’s dispatch practices and provides sufficient information for management to address potential issues or improvements.

MANAGEMENT RESPONSE The Department agreed with the audit recommendations. City Auditor’s Office

City Auditor 480 312-7867 Integrity Line 480 312-8348

www.ScottsdaleAZ.gov

Page 2 Audit Report No. 1703

Patrol Operations Page 3

Center for Public Safety Management (CPSM), formerly within the International City/County Management Association (ICMA) Information and Assistance Division, provides public safety workload and deployment analysis, training and research. CPSM reported it has conducted more than 200 similar studies in 36 states and 155 communities ranging from 8,000 to 800,000 in population.

BACKGROUND

In May 2015, the former City Manager contracted for the Center for Public Safety Management (CPSM) to review operations of the Scottsdale Police Department. CPSM conducted its study during July through December 2015. The final report, delivered in January 2016, included certain operational statistics and compared these results to other police departments that CPSM had reviewed.

In April 2016, in conjunction with the department’s FY 2016/17 budget request, the Police Chief proposed a reorganization plan reducing ten sworn police officer positions and adding 3.6 full-time equivalent positions in other areas, including Records, Communications and parking control. The proposed officer reductions were to be achieved through attrition. The Chief noted the department’s budget request and reorganization plan were partly in response to the CPSM study, as well as addressing other identified organizational resource needs and streamlining for efficiency savings.

At this time, the Council directed that the proposed reorganization be incorporated into the FY 2016/17 budget, but that the added positions not be filled until further review is conducted. To address questions regarding data presented in the CPSM report, the Council directed the charter officers to take “a deeper dive” into the report. In response, the City Treasurer’s office conducted further analysis into the CPSM-reported data.

In May 2016, the Council approved the Acting City Manager’s proposal that the department’s full reorganization be accepted with the added positions, the reorganization be given six months to be implemented, and Patrol Operations would then be subject to audit to evaluate further changes.

The City Auditor included an audit of Patrol Operations in the FY 2016/17 Audit Plan, which was Council-approved.

The Police Department reported that 9 Patrol Officer positions were eliminated by October 2016 through attrition, and an additional 3 positions moved from Patrol to the Bike Unit and 1 position moved to the Training Unit, a total reduction of 13 Patrol Officers, or about 7.3%.

This audit analyzes the department’s FY 2016/17 dispatch data through March 2017 to calculate operational statistics for comparison to the January 2016 CPSM-reported statistics. The CPSM operational statistics are based on analysis of the department’s FY 2014/15 dispatch data.

(continued on next page)

Page 4 Audit Report No. 1703

Figure 1. Patrol District Map

SOURCE: Scottsdale Police Department

Patrol Services

Patrol Services within the Uniformed Services Bureau (USB) of the Police Department is responsible for responding to citizen calls and policing the four geographic districts within the City.

In FY 2016/17, the department deployed 165 police officers and 22 police aides for patrol functions across the 4 designated districts. Patrol squads are groups of 5 to 7 officers, each supervised by a Sergeant. The squads work ten-hour shifts, with some overlapping during busier times of the day. Patrol Services’ total budget of 169 full-time equivalent sworn officers includes four Crime Prevention Officers, one at each district. In addition, 28 sergeants, 10 lieutenants, 4 commanders, 7 administrative police aides and one administrative position were budgeted.

Computer-Aided Dispatch (CAD) data

The department’s CAD system logs Calls for Service events as well as police units’ availability status and location. CAD integrates with the mobile data terminals (MDT) in officers’ vehicles as well as with other police systems, such as the Report Management System.

For 911 calls, call information may be transferred automatically from the phone system. Otherwise, the department’s Call Taker enters the call source and caller information. After obtaining sufficient information from the caller, the Call Taker creates a Call for Service entry and assigns an initial call type and priority number. This information goes into the dispatch queue where the Dispatchers assign available patrol units based on the call’s priority number. The CAD system also shows Dispatchers the location of Patrol units so that the closest may be assigned.

CAD provides a timestamp record of when a call is answered, a Call for Service entry is created, and a police unit is dispatched. It also creates a timestamp record when a police unit reports a change in its status, such as “en route” or “arrived” at the destination.

CPSM and auditors used this CAD data to calculate response and workload statistics.

Dispatch Priorities

The assignment of priority numbers to CFS events is at the Call Taker’s discretion and priorities 0, 1 and 2 are used for high priority calls. Priority 0 calls are officer-initiated priorities, occurring when an officer indicates an emergency need for assistance. Priority 1 is used for calls with immediate threat of danger or loss. For example, “Burglary in Progress,” “Threat with a Gun,” and “Medical Emergency/Not Breathing” were common priority 1 call-types we observed. Priority 2 calls are typically Medical-related Emergencies. Calls are

Patrol Operations Page 5

assigned lower priorities if the immediate threat has passed. The priority assignment and/or call-type is sometimes changed by the Call Taker or Dispatcher, possibly to correct a data entry error, or because the call-type was changed later. For example, calls initially categorized as Priority 1, Medical/Unconscious, were sometimes later changed to Priority 4, Dead Body.

Figure 2. Scottsdale Police Department’s Dispatch Priorities

SOURCE: Auditor analysis of the department’s Dispatch priorities.

Call Sources

Calls for Service (CFS) may be received through various sources but fall into two main categories: citizen-generated and officer-initiated. Citizen-generated CFS’s are those reported by citizens, primarily through the 911-emergency line or the police department’s main phone line. Officer-initiated CFS’s are reported by police officers through their MDT or radio. These CFS’s often include traffic stops and other concerns identified by officers.

Figure 3 on page 6 provides an overview of the Call for Service process.

Page 6 Audit Report No. 1703

Figure 3. Calls for Service (CFS) Process

SOURCE: Auditor analysis of dispatch process and time measures.

Patrol Operations Page 7

OBJECTIVES, SCOPE, AND METHODOLOGY

An audit of Patrol Operations was included on the City Council-approved fiscal year (FY) 2016/17 Audit Plan. The audit objective was to evaluate the effect of the FY 2016/17 staffing reorganization on previous CPSM study results for Patrol Operations, such as comparative workloads, response times and time spent on calls.

To gain an understanding of the organizational changes proposed by the Scottsdale Police Department (SPD), we reviewed the department’s March 2016 memorandum outlining the proposed changes and their budget impacts. We also reviewed the Council meeting minutes where these proposed changes were discussed.

As a baseline for our evaluation, we reviewed the January 2016 Police Operations and Data Analysis Report, Scottsdale, Arizona, a consultant’s study conducted by the Center for Public Safety Management, LLC (CPSM study). From this study, we focused on the following performance statistics that could be impacted by a reduction in Patrol staffing:

• Patrol response times • Time spent on Calls for Service (CFS) • Number of Patrol units responding to CFS • Proportion of Patrol time responding to citizen CFS versus Police-initiated CFS and

other patrol activities.

To gain an understanding of the CPSM study’s methods and deployment practices for Patrol Services, we interviewed the Police Planning Research and Accreditation Director, and the Assistant Chief of Police over the Uniformed Services Bureau. We also contacted CPSM’s Data Team leader for clarifications on the methodology used.

To calculate performance statistics, we requested the data files from the Computer-Aided Dispatch (CAD) system for the current fiscal year, July 1, 2016, through March 23, 2017. This data, totaling more than 2 million records for the 9-month period, included all CFS events and status changes for all police units. To gain an understanding of the data, we interviewed the Sr. Systems Integrator who manages this data. By interviewing the Communications Center Manager, we also clarified our understanding of dispatch procedures to better interpret the data.

To make comparisons to the study’s results, we applied the same methods described in the CPSM report. Where alternative methods seem applicable, we report results based on both methods. To ensure comparability, we also used the same event type classifications as were cited in the CPSM report Appendix.

In addition, when calculating the response-related statistics, we excluded the following irrelevant data records:

• Duplicate entries or unit statuses, • Test entries created by SPD’s IT staff, • CFS events without an officer “Arrive” entry, • CFS events that were not dispatched, • CFS events that were assigned to a non-Patrol primary unit, such as special operations

or traffic enforcement, and • Event types classified as Directed Patrol or non-CFS, such as special details, close

patrol assignments and administrative assignments like mail-runs and training.

Page 8 Audit Report No. 1703

To be consistent with CPSM’s methodology for workload comparisons, we selected two 4-week periods: August 1 through 28, 2016, (summer) and February 19 through March 18, 2017, (after the peak event season). Both of these selected periods occurred after Patrol deployment was reduced to 165 officers. However, in August, there were 2 to 4 additional officers “in-training” to replace the retiring officers, and in the February/March period, there were 4 to 5 vacancies. For this comparison, we analyzed all occupied times from dispatch to clearing the call, regardless of whether the officer logged an arrival time.

The Patrol position reductions have had limited effect on the operational measures that were previously reported by CPSM. There were no significant changes in emergency response times, slightly higher average occupied time, and slightly fewer Patrol units responding to calls than previously reported. With 13 fewer positions, Patrol is spending a little more time on reactive calls, but the specialty units augment Patrol’s proactive time. Additionally, procedural changes regarding system testing and response time reporting can better assure reliable data for operational analysis.

We conducted this audit in accordance with generally accepted government auditing standards as required by Article III, Scottsdale Revised Code §2-117 et seq. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. Audit work took place from March to May 2017.

Patrol Operations Page 9

Time Measures

Dispatch – time from Call Taker’s CAD entry to the Dispatcher assigning a Patrol Unit.

Travel – time from call dispatch to the Patrol unit’s arrival on scene.

Response – Dispatch and Travel time.

FINDINGS AND ANALYSIS

1. The Patrol position reductions have had limited effect on the operational measures that were previously reported by CPSM.

Patrol position reductions have not resulted in significant changes in response times.

With the reduction of 13 sworn Patrol positions since July 1, 2016, there has not been a significant increase in response times. Using CPSM’s analytical methodology, average response times for high priority Calls for Service increased only slightly, from 6.10 minutes to 6.18 minutes, or by about 1.3%.

Table 1 on page 10 summarizes the comparative response times, detailing the Dispatch, Travel, and total Response time. These times are recorded in the Computer-Aided Dispatch (CAD) system based on timestamps for call taker, dispatcher and police officer system entries.

Dispatch times can be affected by the availability of Patrol units and severity of the call, or its priority, relative to other calls waiting in the queue to be dispatched.

Priority 0, 1 and 2 are defined by the department as emergency calls to be dispatched immediately before others in the queue.

(continued on next page)

Page 10 Audit Report No. 1703

Table 1. Average Response Times in Minutes: Citizen-generated Calls for Service

FY 2014/15 (CPSM) July 2016 – March 2017 (Audit)

Priority 1 Dispatch Travel Response 2 Dispatch Travel Response 2

1 1.1 4.4 5.6 0.9 4.8 5.6 2 1.2 5.7 6.8 1.5 5.8 7.3 3 1.8 6.0 7.8 2.5 6.7 9.2 4 4.3 7.4 11.7 4.7 7.7 12.4 5 5.7 9.1 14.8 6.1 9.7 15.8 6 11.6 10.0 21.6 7.7 13.6 21.3 7 24.4 14.4 38.8 22.8 21.5 44.3 8 26.3 11.9 38.2 34.4 16.1 50.5 9 29.7 13.1 42.9 26.7 13.5 40.2

High-Priority Average 1.14 4.94 6.10 1.08 5.10 6.18

1 – There were no “0” priority calls. Auditor analysis based on initial Priority number assigned by the Call Taker. This initial dispatch priority may change as additional information becomes available during the call. CPSM did not answer our inquiry of whether the initial or final priority was used in its analysis. 2 – Response time and High Priority average are averages of individual call totals, rather than the sums of other categories.

SOURCE: Data table 7-19 from page 174 of the CPSM’s January 2016 report and auditor analysis of CAD data for July 1, 2016 through March 23, 2017.

• “Response” time in Table 1 does not completely align with the department’s response time goal definition. The department’s performance goal is to respond to emergency calls (those categorized as 0, 1 or 2) within 5 minutes or less. The department frequently describes this time measure as “Hello to Hello,” or from the time the call is received to the time the officer arrives on-scene. However, CPSM’s data analysis did not include Call Taker time, and therefore, our FY 2016/17 audit analysis also does not. So response times shown in Table 1 are understated by the Call Taker time, which elapses from receiving the call to creating the CAD entry to be dispatched.

For FY 2014/15, the department reported an average “Hello to Hello” emergency response time of 5.42 minutes, lower than CPSM’s calculated 6.1 minutes without the Call Taker time. As detailed in Finding 5(B), the department’s methodology for monitoring response times excludes certain calls and the report criteria should be reviewed to ensure the department is monitoring what is intended.

• As well, CPSM did not include in its analysis any Call for Service events with dispatch delays of more than 3 hours, assuming these entries may have been erroneous. However, SPD defines priorities 7 through 9 as “Routine” calls that are to be dispatched to Patrol units as time permits or, when applicable, at a time and date the caller requested. When we also include calls with 3-plus hours dispatch delay, these low priority calls (7 through 9) averaged 40 to 50 minute dispatch times and 54 to 72 minute response times.

Patrol Operations Page 11

Average occupied time is now slightly higher than in the CPSM study.

“Occupied time” refers to the amount of time a patrol officer spends on a Call for Service. For citizen-generated calls, this measures the time from when the officer is dispatched until the officer records clearing the call. For officer-generated calls, this time is measured from when the officer notifies dispatch of beginning the activity until reporting it is cleared.

Auditors analyzed the primary unit assigned to the call, consistent with CPSM’s method. Similarly, the analysis also excluded events where the officer is recorded as spending less than 30 seconds at the scene or more than 8 hours.

As shown in Table 2, the overall average occupied time for citizen-generated Calls for Service increased from about 43.8 minutes to about 49 minutes, an increase of 11.8%. For citizen-generated calls, the largest increases occurred in Juvenile crimes, Alarm calls, Traffic Enforcement, and Property crimes, while average time per call spent on Medical calls decreased. The overall average time spent on officer-initiated calls increased only slightly from 27.9 to 28.6 minutes, or 2.4%.

Table 2. Average Occupied Time per Call in Minutes, by Call Type and Source

FY 2014/15 (CPSM) July 2016 - March 2017 (Audit)

Citizen-Generated

Officer-Initiated

Citizen-Generated

Officer-Initiated

Accidents 75.4 62.7 76.9 61.3 Alarms 17.2 13.2 20.0 8.2 Animal Calls 28.2 27.7 30.8 20.9 Assist Other Agency 48.7 27.0 51.9 33.1 Check/Investigation 30.6 15.3 33.4 14.0 Citizen Assist 47.5 31.8 51.1 23.4 Crime-Drug 132.1 108.7 140.8 128.0 Crime-Person 84.1 92.5 89.9 105.9 Crime-Property 74.9 73.0 83.6 84.7 Disturbance 41.5 72.2 43.2 88.9 Follow up 56.4 70.4 61.2 71.6 Juvenile 51.4 55.4 61.7 62.9 Medical 35.4 32.3 30.8 20.0 Miscellaneous 50.7 52.9 51.5 56.1 Prisoner-Arrest 108.3 98.5 106.9 104.6 Prisoner-Transport 87.2 84.2 89.7 88.6 Traffic Enforcement 23.4 15.6 27.0 15.2

Average Time 43.8 27.9 49.0 28.6 SOURCE: Table 7-6 on page 138 of CPSM’s January 2016 report and auditor analysis of CAD data for July 1, 2016, through March 23, 2017.

Page 12 Audit Report No. 1703

During data analysis, we noted that the primary unit assigned in CAD was sometimes not the one spending the most time on a particular call. Also, some units returned to the scene multiple times and sometimes over multiple days. For these instances, the largest time differences were for crime-related calls (Drug and Persons) and prisoner arrest. For example, citizen-generated calls for Drug-related crimes averaged 169 minutes for the longest occupied unit. The overall average occupied time increased slightly with this method, from 49 minutes to 50.2 minutes.

After the 13-position reduction, Patrol’s average occupied time of 49 minutes remains longer than the 28.7 minute average that CPSM reported was observed in its past reviews of 56 other municipal police departments. Because these municipalities varied greatly in geographic size, population and crime rates from Scottsdale, we researched whether other comparison data was available. However, we did not find any published public safety standards or benchmarks for time spent on calls for service. Other Valley cities also do not publish this type of performance measure that we could use for comparison purposes.

On average, fewer Patrol units are now responding to calls than previously reported.

With the reduction of 13 sworn Patrol positions, fewer units responded to calls in the FY 2016/17 period reviewed than reported in the CPSM study.

Dispatchers assign patrol units to Calls for Service based on the call type and the department’s guidelines. Table 3 on page 13 summarizes the number of Patrol units that responded to Calls for Service and arrived at the scene.

(continued on next page)

Patrol Operations Page 13

Table 3. Average Number of Responding Patrol Units, by Call Type and Source

Citizen-Generated Officer-Initiated

Category FY 2014/15

(CPSM) 2016/17 (Audit)

FY 2014/15 (CPSM)

2016/17 (Audit)

Accidents 2.2 2.0 2.0 2.1 Alarms 1.7 1.4 1.5 1.1 Animal Calls 1.2 1.1 1.3 1.2 Assist Other Agency 2.1 2.0 1.3 1.3 Check/Investigation 2.1 1.8 1.5 1.4 Citizen Assist 2.0 1.8 1.4 1.2 Crime-Drug 2.8 2.7 2.4 2.5 Crime-Person 2.1 2.5 2.1 2.1 Crime-Property 1.6 1.6 1.6 1.7 Disturbance 2.4 2.2 2.3 2.4 Followup 1.2 1.1 1.1 1.1 Juvenile 2.2 1.9 1.7 1.4 Medical 2.2 1.8 1.6 1.6 Misc. 1.3 1.2 1.5 1.4 Prisoner-Arrest 2.4 2.4 2.3 2.2 Prisoner-Transport 1.2 1.1 1.3 1.3 Traffic Enforcement 1.5 1.2 1.3 1.3

Average Number of Units 1.9 1.7 1.4 1.3

Analysis includes Patrol’s responding Police Aide units but excludes non-patrol units, such as crime scene specialists, detectives, or tactical units.

SOURCE: Table 7-7 on page 140 of CPSM’s January 2016 report and auditor analysis of CAD data from July 1, 2016, through March 23, 2017.

Compared to the previous CPSM-reported results, in FY 2016/17 fewer patrol units are being dispatched for citizen-generated calls in all categories except for Persons-related crimes. The department has been dispatching 2 or fewer units for about 87 percent of calls, as shown in Figure 4 on page 14.

(continued on next page)

Page 14 Audit Report No. 1703

Figure 4. Number of Responding Patrol Units, by Percentage of Calls For Service

SOURCE: Table 7-8 on page 142 of CPSM’s January 2016 report and auditor analysis of CAD data for July 1, 2016, through March 23, 2017.

The current fiscal year’s overall average of 1.7 units dispatched for citizen-generated Calls For Service is very similar to the 1.6 average units reported by CPSM as observed in its past reviews of other municipal police departments. However, as with the Occupied Time measure, we did not find public safety standards or benchmarks for the number of responding units, nor similar statistics reported by other Valley cities for comparison.

For the current fiscal year, we also reviewed trends in the number of responding units by the time of day. As shown in Figure 5 on page 15, there is a notable increase in the number of patrol units dispatched in the evening and early morning hours for some call types, such as Persons crimes, Property crimes, Alarms and Accidents.

(continued on next page)

42.2%

34.4%

23.4%

52.4%

34.3%

13.3%

0%

10%

20%

30%

40%

50%

60%

One Two Three or More

Citizen-generated Calls for Service

FY 2014/15 (CPSM) FYTD 2016/17

Patrol Operations Page 15

Figure 5. Average Number of Responding Patrol Units, by Call Type

Based on observed patterns of natural breaks in the data, we grouped Calls for Service that occurred between 7 pm and 6 am as “Evening/Early Morning.”

SOURCE: Auditor analysis of CAD data for July 1, 2016, through March 23, 2017.

(continued on next page)

0 1 2 3

Prisoner-Transport

Followup

Animal Calls

Misc.

Traffic Enforcement

Citizen Assist

Alarm

Check/ Investigation

Medical

Juvenile

Crime-Property

Assist Other Agency

Disturbance

Prisoner-Arrest

Accident

Crime-Drug

Crime-Person

Average No. of Patrol Units

Citizen-generated CFS

Day Evening/Early Morning

0 1 2 3

Alarm

Followup

Animal Calls

Assist Other Agency

Citizen Assist

Traffic Enforcement

Check/ Investigation

Prisoner-Transport

Medical

Misc.

Juvenile

Prisoner-Arrest

Disturbance

Crime-Property

Crime-Drug

Accidents

Crime-Persons

Average No. of Patrol Units

Officer-initiated CFS

Day Evening/EarlyMorning

Page 16 Audit Report No. 1703

Patrol is spending more time on reactive calls, but specialty units augment Patrol’s proactive time.

With 13 fewer positions, Patrol units are spending more time on reactive (citizen-generated) calls. However, other Uniformed Services specialty units, such as Bike, Motor and DUI, supplement Patrol’s reduced proactive time.

Based on the CPSM study, Patrol workload can be analyzed in the following categories:

• Reactive – citizen requests for assistance • Proactive: Police-initiated - officer-initiated CFS, close patrol, community policing

activities, directed activities, or patrol details • Proactive: Uncommitted - patrolling but not performing a specific activity • Administrative - breaks, meals, training, court attendance, equipment check,

fueling, meetings, paper-work

In the past, the department’s staffing model was based on 26.4% reactive, 33.6% proactive, 20% uncommitted, and 20% administrative. Its current deployment model specifies at least 20% proactive time per hour, but does not establish percentages for the other workload categories.

Figure 6 on page 17, illustrates the percentage of time spent for each of these categories compared to available time for standard Patrol deployments. Compared to CPSM’s reported FY 2014/15 results, Patrol officers are spending slightly more time responding to citizen calls and Police-initiated events, and had less Uncommitted time.

Similar to the CPSM study method, auditors used CAD data from a 4-week summer period (August 2016) and a 4-week winter period (February/March 2017) to calculate time spent in these four categories. However, CPSM did not provide further clarification on the methodology used for this calculation. So, for the audit analysis, we calculated the total available hours based on the police officer and police aide positions that are assigned to the patrol function. The Patrol sergeants, lieutenants and commanders are primarily supervisory positions, and their available hours are not included.

We also reduced these total hours by the 15.9% leave factor that the department uses in its deployment model. This factor represents the average amount of leave hours (vacation, medical, and other types) taken by the patrol positions during a fiscal year.

To ensure comparable results are presented, we used the same method to recalculate FY 2014/15 results. These adjusted FY 2014/15 results are presented in Figure 6 on page 17 as “CPSM Adj’d.”

(continued on next page)

Patrol Operations Page 17

Figure 6. Patrol Time Spent on Calls for Service and Other Activities in Selected Periods

Reactive and Proactive: Police-Initiated are based on calls for service recorded in CAD. Admin is based on officer-reported “out-of-service” categories as recorded in CAD. However, out-of-service status recorded as “detail” is included in the Proactive: Police-Initiated category. Proactive: Uncommitted is the remaining available time that was not recorded as specific activities.

SOURCE: Auditor analysis averaging CAD data for August 1 to August 28, 2016, and February 19 to March 18, 2017, and Table 6-1 on page 81 of the CPSM’s January 2016 report, which averages CAD data for August 1 to August 28, 2014, and February 1 to February 28, 2015. “CPSM Adj’d” is based on auditor analysis of CAD data for the same August 2014 and February 2015 periods, adjusted for the additional police aide positions and the leave factor.

Accuracy of this time analysis depends on the officers consistently reporting their activities. For example, the department classifies writing reports for Calls for Service as administrative time. However, if an officer writes reports between calls but remains “available” (does not log activity in CAD), the time will be categorized as Uncommitted rather than Administrative.

Specialty units augment Patrol’s proactive policing activities

Proactive policing is police-initiated activities to deter crime by maintaining a visible presence in the community and interacting with the public. Examples include Patrol officers driving through neighborhood streets periodically, making traffic stops, or being assigned to identified problem areas. Proactive policing sometimes results in Officer-initiated Calls for Service, when the officer identifies a potential crime.

Figure 6 shows that Patrol more than doubled the department’s 20% target for proactive activity in both fiscal year periods. However, the department’s proactive policing activities are greater than the Patrol workload alone demonstrates.

26.8% 29.7% 31.4%

16.5% 19.1% 21.3%

48.2% 43.8% 39.0%

8.5% 7.4% 8.3%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

CPSMFY14/15

FY14/15 -CPSM Adj'd

FY16/17 -Audit

Admin

Proactive:Uncommitted

Proactive:Police-Initiated

Reactive

Page 18 Audit Report No. 1703

Specialty Units

These units are not included in Patrol analysis:

Bike – ten officers on bicycles primarily patrolling District 2 (downtown).

Motor – nine officers on motorcycles focused on traffic enforcement.

DUI – six officers focused on DUI patrols.

Specifically, the Bike and Traffic Enforcement (Motor) Units also perform proactive policing by patrolling assigned areas. Similarly, the DUI Unit assists Patrol when DUI’s are involved, which decreases the time a Patrol unit is occupied on these cases. These units may also be assigned to special events or focus on specific locations and issues. The department does not include these units in the Patrol deployment model because their schedules and assignment areas are subject to change as needed rather than constant coverage.

Based on CAD entries during the 4-week periods in August 2016 and February/March 2017, the Bike and Traffic units responded to citizen Calls for Service about 8 percent of their available time and initiated calls about 22 percent of the time. Uncommitted patrol time was approximately 63 percent.

When these units are included in the analysis, the department’s overall time spent on proactive policing increases by 2.3 percent.

Recommendation: None.

2. Procedural changes can better assure reliable data for operational analysis.

While analyzing the department’s CAD data and response time reporting, we noted areas where procedures can be improved.

A. Test entries used to validate system processing are made in the live (production) CAD system. Therefore, to preserve CAD data integrity, a consistent method of tracking test entries is needed so that they can be removed after testing is completed.

In the nine-month CAD data extract for FY 2016/17, we identified almost 1,200 event numbers that appeared to be test events, as well as some additional test entries that were not associated with event numbers.

The information technology support staff creates test entries to validate the data flows or to test a specific software change. According to the CAD systems integrator, these entries are typically identified as “TEST” in the event code or by an “ADMIN” unit number instead of a typical police unit number. While the majority of the test entries were labeled this way, a small number of entries used fake unit numbers rather than ADMIN. Also, at least 116 events appeared to contain valid data but had “TEST” in the initial event type. These events did not also have TEST in the final event type or incident code where it would normally be identified. Additionally, one apparently valid Call for Service included some test entries within its event history, potentially altering the reliability of that record.

We excluded all identified test entries from our analysis, but it is not clear whether CPSM could have identified all test entries. CPSM appeared to use the final event codes, and we found about 130 test events in the FY 2014/15 data that were only identified in the initial code.

Patrol Operations Page 19

Due to system complexity, the department does not have a duplicate test environment and tests in the live production system. When this approach is determined to be necessary, specific procedures are needed to safeguard the system, including parameters to identify test entries to allow their removal. Further, system procedures should specify appropriate documentation and approval of any data deletion, including test entries.

We did not perform a comprehensive review of information system controls during this audit, so other improvements may be needed.

B. The department’s internal response time reports exclude certain call types or events, which can distort the results.

• While medical calls are always transferred to the Regional Dispatch Center for Fire Department dispatch, the Police Department also dispatches an officer in certain instances. Particularly for calls reporting someone not breathing, drowning or unconscious, Patrol units are often dispatched as priority 1 in case officers may arrive more quickly to administer CPR. Patrol units may be dispatched on other medical calls at the dispatcher’s discretion based on the circumstances of the call or at the Fire Department’s request for support.

However, the Police Department’s response time reports exclude all medical call types except “Not Breathing” and “Drowning.” For the FY 2016/17 review period, of the approximately 1,800 calls initially classified as high-priority, more than half were Medical-Assist and Medical-Unconscious. Yet these medical-related calls would not have been included in the department’s internally-calculated response time measures. This exclusion can have a significant effect on calculated response times.

Because the Fire Department’s request for support may come after the firetruck has arrived, Medical-Assist Calls reflect a delayed Patrol dispatch time. The average response time for all high priority calls was 6.18 minutes, as shown in Table 1 on page 9. After excluding Medical-Assist call types, the average response time was about 5.6 minutes. When excluding all Medical calls except Drowning and Not Breathing, as the department’s internal report does, the overall average response time was about 4.9 minutes.

• Additionally, the department’s response time reports do not capture SPD’s emergency response when the priority number has been changed. For example, a call that started as priority 1 “Shooting” but later changed to a priority 4 “Dead Body” is excluded from the analysis even though it is not clear whether the classification was changed before or after police arrived. Conversely, it also does not capture events where a low priority call becomes high-priority, such as a “Welfare check” call that becomes priority 0, “Officer Involved Shooting.” Excluding calls that changed priority significantly reduces the number of events reviewed. Approximately 29 percent of the calls that were initially classified as high priority were changed to a different priority number.

Besides providing the department’s calculated average response times, these reports are used by the Communications Center to review call taker and dispatcher performance. However, because the report excludes atypical high-priority calls, department management does not see these and cannot address potential issues.

Page 20 Audit Report No. 1703

Recommendations:

The Police Chief should:

A. Direct the department’s Technology group to develop a detailed procedure for any testing conducted in a live system. The procedure should require test entries to be consistently identified, monitored and removed, if possible. In addition, the procedure should specify documentation and approval for any data modification or deletion.

B. Ensure that the internal response time report is consistent with the department’s dispatch practices and provides sufficient information for department management to address potential issues or improvements.

Patrol Operations Page 21

MANAGEMENT ACTION PLAN

1. The Patrol position reductions have had limited effect on the operational measures previously reported by CPSM.

Recommendation: None

2. Procedural changes can better assure reliable data for operational analysis.

Recommendations:

The Police Chief should:

A. Direct the department’s Technology group to develop a detailed procedure for any testing conducted in a live system. The procedure should require test entries to be consistently identified, monitored and removed, if possible. In addition, the procedure should specify documentation and approval for any data modification or deletion.

B. Ensure that the internal response time report is consistent with the department’s dispatch practices and provides sufficient information for department management to address potential issues or improvements.

MANAGEMENT RESPONSE: Agree

PROPOSED RESOLUTION:

A. The Department will evaluate the process in which testing is conducted in the live system and document the appropriate procedure in a published policy.

Upon creation of this procedure, department staff in Technology Services and Communications will receive training on this procedure to ensure consistency in use. Further, we will ensure that all departmental reports used for querying calls for service, and subsequent statistics, will exclude any calls for service generated with “test” parameters.

B. The Department will evaluate all calls for service and associated reports.

Staff will conduct a full review of call types and priorities and ensure all are properly categorized. A corresponding policy will be published that ensures consistency and clarification for all call types. In addition, the Department will complete a full review of all relevant reports to ensure parameters (including titles, disclaimers, queried data, locations, etc.) are accurate representations of the information we are seeking to provide.

To ensure that all valid calls are analyzed, staff will create a report to identify calls for service in which a change in priority occurred mid-call. This will be available for the communication manager to run and distribute to appropriate department staff when there are concerns about response times.

RESPONSIBLE PARTY: Alan Rodbell, Chief of Police

COMPLETED BY: 6/14/2017

City Auditor’s Office 7447 E. Indian School Rd., Suite 205 Scottsdale, Arizona 85251 OFFICE (480) 312-7756 INTEGRITY LINE (480) 312-8348 www.ScottsdaleAZ.gov/auditor

The City Auditor’s Office conducts audits to promote operational efficiency, effectiveness, accountability, and integrity.

Audit Committee Vice Mayor Suzanne Klapp, Chair Councilmember Virginia Korte Councilwoman Kathy Littlefield City Auditor’s Office Kyla Anderson, Senior Auditor Lai Cluff, Senior Auditor Cathleen Davis, Senior Auditor Brad Hubert, Internal Auditor Dan Spencer, Senior Auditor Sharron Walker, City Auditor